Skip to content
Online marketing in the agent era
ROA·Marketing
Menu

Competitive Intelligence

competitor-analysis-v2

Structured competitive intelligence — scrape + screenshot client and ~5 competitors, extract messaging/positioning, run gap analysis, verify opportunities against the client's actual capabilities, deliver a client-facing strategy PDF ("Client Gift") and/or a 1-page tactical Ads Angle Brief

Trigger phrasesanalyze competitorscompetitor reportcompetitive intelligenceclient giftmarket positioning analysis
Read-only Part of a 43-skill library Updated 2026-07-27 · MIT Companion skills: 2 📜 Python scripts

Install this skill

bash
mkdir -p .claude/skills/competitor-analysis-v2 && curl -o .claude/skills/competitor-analysis-v2/SKILL.md https://roa-marketing.com/skills/competitor-analysis-v2/SKILL.md
On this page
  1. What this skill does
  2. Requirements
  3. Using it
  4. Inputs
  5. The 8 phases
  6. Quality gate before delivery
  7. The skill file
  8. FAQ

What this skill does

competitor-analysis-v2 is a Google Ads competitive intelligence skill that analyzes your account data. Install it as a Claude Code skill and invoke it with its trigger phrases to automate a recurring PPC workflow.

  • Automates a competitive intelligence task that would otherwise require manual work in the Google Ads interface
  • Ships with Python scripts that connect directly to the Google Ads API for production use
  • Integrates with the broader ROA Marketing skills library — pairs with companion skills for end-to-end workflows

Requirements

✓ Claude Code installed
✓ Google Ads API access configured
✓ Python with: google-ads

Using it

Illustrative example

Youcompetitor-analysis-v2
Agent
Running competitor-analysis-v2 on your account…

This is a read-only skill — no changes will be made to your account.

Processing complete. See the output above for results specific to your account configuration.

The skill file

Everything below is the instruction set your AI agent follows — you install it, the agent reads it.

Two deliverables from one analysis: Client Gift (10–15pp strategy PDF for the client — zero Google Ads mechanics) and/or Ads Angle Brief (1-page tactical doc for the practitioner). Ask which at the start; "both" is the full-service default. Public data only — competitor websites and visible ads; nothing requiring access you don't have.

Inputs

Client name + URL · industry/market · ~5 competitor URLs (none provided → WebSearch 8–10 candidates, present with one-liners, user picks 5) · output choice. If multiple credential profiles exist in your setup, lock one before any Ads-side queries (Ads data is optional — website-only mode is fully supported).

The 8 phases

  1. Inputs (above).
  2. Screenshots (mandatory): node scripts/screenshot.cjs <client> <c1> … <c5> --output /tmp/competitor-report/screenshots (Playwright) → 6 full-page PNGs + metadata. Visual evidence makes the deliverable.
  3. Parallel extraction: 6 concurrent agents, one per site, each running the prompts/website-extraction.md prompt via WebFetch → structured profile (value prop, services, trust signals, social proof, guarantees, emotional triggers, audience, tier, tone). Parallel ≈ 1 min vs ~6 serial.
  4. Ad analysis (if competitor ads available via SERP capture or paste): strategic 15-attribute + tactical 7-attribute scoring (prompts/). No ads → skip and note it; website intelligence sustains the Client Gift alone.
  5. Gap identification: messaging matrix (competitor × headline/claim/USP/CTA/proof/pricing) → claims in 3+ sites = table stakes; claims in 1 = differentiators; claims in 0 = whitespace. Positioning map (2×2, axes fit to vertical: price↔specialization, speed↔quality, scale↔personal). Categorize: strategic gaps / tactical gaps / unmet needs.
  6. Client verification (mandatory, non-negotiable): scrape the client site; map each gap → ✅ can fill (site evidence, cited) or ❌ cannot verify (goes on the explicit cannot-recommend list). This is ad-copy-verification-standard applied to strategy — no recommending positions the client can't substantiate.
  7. Generate: Client Gift per templates/client-gift.md (exec summary → competitor profiles → screenshots grid → matrix → table stakes → opportunities → positioning map → implications → recommendations) — strategic tone, no platform mechanics. Ads Brief per templates/ads-angle-brief.md: top-3 verified angles + proof, angles not to pursue, headlines per angle, campaign→angle map, compliance watch-outs (regulated verticals: run fair-housing-compliance or equivalent).
  8. PDF (Client Gift): HTML + templates/template.css → Playwright page.pdf() (A4, backgrounds on).

Quality gate before delivery

6/6 screenshots and extractions · Phase 6 done with citations on every recommendation · cannot-recommend list present · no Ads mechanics in the Client Gift · Ads Brief uses verified angles only.

FAQ

Does this skill change my Google Ads account?

No — this is a read-only skill. It queries the Google Ads API for reporting and analysis, but makes zero changes to campaigns, keywords, budgets, or settings.

What does it cost to run?

The skill itself is free and MIT-licensed. Google Ads API usage may incur standard Google API quota costs depending on your access tier — most standard accounts stay well within free quota limits for reporting operations.

Which accounts does it work on?

Any Google Ads account with API access configured. Single-account and MCC-level skills are labeled in their descriptions — check the skill page for account scope details.

Don't want to run this yourself?

We run this audit (and 42 more) on client accounts — Talk to us.

Based on ppc-ai-skills by Kurt Henninger (MIT), analyzed and rewritten by ROA Marketing.